Digital twin simulation modeling and online optimization of the mandrel feed strategy for profiled ring rolling

The mandrel feed strategy in profiled ring rolling (PRR) determines the deformation path of the ring and strongly affects both geometric accuracy (roundness) and microstructure (grain size and grain uniformity). However, due to the strong nonlinearity and path dependency of PRR, the evolution of roundness, grain size, and grain uniformity cannot be explicitly modeled. In addition, the complex evolution of ring states leads to stage-dependent control objectives during PRR. These factors make existing predefined mandrel feed strategies ineffective for achieving coordinated control of ring geometry and microstructure. To address this challenge, an artificial neural network-based model predictive control (ANN-MPC) framework for PRR is proposed. The framework predicts the ring’s geometric and microstructural states over a finite horizon and optimizes the mandrel feed rate according to predefined objectives. Data-driven state equations are constructed using ANNs with embedded physics constraints, ensuring physically consistent state prediction and control decisions. Considering the path dependency of PRR, a cost function with state-dependent weights is designed. This allows staged priority scheduling among control objectives according to the evolving state risks, thereby achieving coordinated control of roundness, grain size and grain uniformity of the rolled ring. Based on this framework, a digital twin finite element model (DTFEM) is developed for PRR. It continuously monitors, analyzes, and predicts the ring states, and optimizes and executes the mandrel feed strategy online, achieving coordinated control of ring geometry and microstructure within a single simulation. This work provides crucial theoretical and technical support for the research and development of AI-driven ring rolling technology and intelligent ring-rolling equipment.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-20
DOI
https://doi.org/10.21203/rs.3.rs-11003106/v1
Primary Topic
Metal Forming Simulation Techniques
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Digital twin simulation modeling and online optimization of the mandrel feed strategy for profiled ring rolling

Luo Hengjun, Hao Deng, Lianggang Guo, Heng Li et al.
2 links
Research Square
Metal Forming Simulation Techniques
preprint

Digital twin simulation modeling and online optimization of the mandrel feed strategy for profiled ring rolling

Luo Hengjun, Hao Deng, Lianggang Guo, Heng Li, Haotian Li
preprint en

Abstract

The mandrel feed strategy in profiled ring rolling (PRR) determines the deformation path of the ring and strongly affects both geometric accuracy (roundness) and microstructure (grain size and grain uniformity). However, due to the strong nonlinearity and path dependency of PRR, the evolution of roundness, grain size, and grain uniformity cannot be explicitly modeled. In addition, the complex evolution of ring states leads to stage-dependent control objectives during PRR. These factors make existing predefined mandrel feed strategies ineffective for achieving coordinated control of ring geometry and microstructure. To address this challenge, an artificial neural network-based model predictive control (ANN-MPC) framework for PRR is proposed. The framework predicts the ring’s geometric and microstructural states over a finite horizon and optimizes the mandrel feed rate according to predefined objectives. Data-driven state equations are constructed using ANNs with embedded physics constraints, ensuring physically consistent state prediction and control decisions. Considering the path dependency of PRR, a cost function with state-dependent weights is designed. This allows staged priority scheduling among control objectives according to the evolving state risks, thereby achieving coordinated control of roundness, grain size and grain uniformity of the rolled ring. Based on this framework, a digital twin finite element model (DTFEM) is developed for PRR. It continuously monitors, analyzes, and predicts the ring states, and optimizes and executes the mandrel feed strategy online, achieving coordinated control of ring geometry and microstructure within a single simulation. This work provides crucial theoretical and technical support for the research and development of AI-driven ring rolling technology and intelligent ring-rolling equipment.

Research Square
Northwestern Polytechnical University (CN), State Key Laboratory of Solidification Processing (CN)
Industry, innovation and infrastructure
Metal Forming Simulation Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon